#is there a way to obtain the KB RAG search query in custom llm server
1 messages · Page 1 of 1 (latest)
is there a way to obtain the KB RAG search query in custom llm server
Short answer: No. The platform performs RAG (including any query rewriting) inside ElevenLabs before calling your Custom LLM Server, and it does not expose the RAG search query in the OpenAI‑compatible chat.completions request your server receives. The request contains the normal messages (with retrieved KB context already injected), not the raw retrieval query. (elevenlabs.io)
What you can access instead
- Retrieved KB chunks for a conversation: available in the dashboard history and via the Conversations API; ElevenLabs added RAG chunk details to conversation history and to the Get Conversation response. (elevenlabs.io)
- The content of specific chunks: use Get knowledge base document chunk to fetch the text for any chunk referenced in the conversation. (elevenlabs.io)
Notes
- Custom LLM Server integration follows the OpenAI chat.completions shape; enabling “Custom LLM extra body” only lets you pass your own extra parameters, not RAG internals. (elevenlabs.io)
If you need the actual RAG search query exposed, that isn’t documented today—please file a feature request with Support. (help.elevenlabs.io)
Sources:
Retrieval-Augmented Generation | ElevenLabs Documentation
March 17, 2025 | ElevenLabs Documentation
Get document chunk | ElevenLabs Documentation
Integrate your own model | ElevenLabs Documentation
How can I contact support? – ElevenLabs
is there a way to obtain the KB RAG search query in the custom llm server